What data sources can Genie One connect to?

Summary

• Genie is an AI agent for business intelligence that accesses data governed in Unity Catalog, making it possible to query any data you've ingested into Databricks.
• The Databricks lakehouse ingests data from hundreds of sources: data warehouses, data lakes, SaaS applications, databases, APIs, and streaming platforms.
• Genie uses SQL and semantic understanding to answer business questions against this data, supporting natural language queries across tables, dashboards, and metadata.
• Data sources are governed through Unity Catalog, so Genie respects role-based access control, ensuring users only see data they're authorized to access.
• Once your data is in Databricks, Genie provides an interface for business users to explore and analyze it without writing SQL or building custom reports.

What data sources can Genie One connect to?
Genie, the AI agent for business intelligence in Databricks, connects to any data you've brought into your Databricks workspace via Unity Catalog. It doesn't directly connect to external data sources, but rather queries your unified data lakehouse. This means Genie's data source reach is determined by the breadth of data you ingest into Databricks.

How Genie accesses data

Genie runs natural language queries against SQL tables, files, and metadata stored in your Databricks lakehouse. It uses semantic understanding of your data (column names, table relationships, business metrics, and lineage) to translate a business question into SQL, execute that SQL, and return results. Because all access goes through Unity Catalog, Genie respects your governance policies: users only see data they have permission to access, and all queries are audited.
The data Genie can query depends entirely on what you've ingested into Databricks. The lakehouse supports ingestion from a broad range of sources through Databricks connectors, Lakeflow Pipelines, and third-party tools.

Common data sources supported

Source category	What it includes
Data warehouses	Proprietary cloud data warehouses and legacy data warehouse systems
Data lakes	Cloud storage and open data lake formats
Databases	Relational databases, NoSQL systems, and transactional platforms
SaaS applications	CRMs, marketing platforms, HR systems, accounting tools, and business applications
Analytics platforms	Dashboarding, BI, and business intelligence tools
Cloud storage	Object storage and file-based data repositories
Streaming systems	Event streaming platforms and real-time data pipelines
APIs and webhooks	REST endpoints, GraphQL, and custom integrations
Files and documents	Spreadsheets, documents, PDFs, and other unstructured formats

How to make data available to Genie

Ingest through Lakeflow. Use declarative data integration with Lakeflow to set up recurring ingestion from data warehouses, databases, and SaaS platforms. Lakeflow handles schema evolution, incremental updates, and data quality checks. Data lands in bronze (raw), silver (cleaned), and gold (business-ready) tables, all governed by Unity Catalog.
Integrate through partner tools. Use third-party data integration platforms and partner connectors to ingest data from various sources. These tools can run ELT logic and load results directly into your lakehouse.
Build custom connectors. If a standard connector doesn't exist, Databricks provides APIs and SDKs (Python, Scala, SQL) to build custom ingestion logic. Run these connectors as scheduled Lakeflow Jobs or real-time streaming jobs.
Real-time streaming. Use Structured Streaming to ingest from streaming platforms, transforming events on arrival and writing results to lakehouse tables that Genie can query.

Governance and access control

Once data is in Databricks, Unity Catalog provides unified governance across all data sources. You can apply column-level and row-level security policies, classify sensitive data, and enforce purpose-based restrictions. Genie respects all these policies: a user's natural language query only accesses tables and columns they have permission to see.
Lineage tracking in Unity Catalog shows where each table came from (source system, last refresh date, data quality checks passed), so users and auditors understand data provenance.

Genie for cross-source analysis

Once multiple data sources are in the lakehouse, Genie can run queries that join data across them. A user might ask, "Show me revenue by customer segment, comparing quarterly actual results against our forecast." Genie joins data from your sales system, forecasting platform, and customer database, answering the question in natural language.

FAQs

Does Genie directly connect to my data sources?

No. Genie queries data already ingested into your Databricks lakehouse through Unity Catalog. Data must be copied or replicated into Databricks first. This architecture provides performance benefits, unified governance, and consistency across all analysis.

How fresh is the data Genie sees?

Data freshness depends on your ingestion schedule. Lakeflow supports batch ingestion (hourly, daily, on-demand) and real-time streaming. Set refresh frequency based on your business needs and source system constraints.

What if my data source isn't listed?

You can build custom ingestion using Python, Scala, or SQL. Lakeflow Jobs can run your custom code on a schedule. Alternatively, export data as CSV, JSON, or Parquet files and upload them to your lakehouse.

How does Genie know which tables to query?

Genie has access to all tables in catalogs and schemas you assign to it. Use Unity Catalog to grant Genie read permissions, then ensure tables are registered in the catalog with meaningful names and descriptions. Genie uses this metadata to understand what data it can access.